It seems that we are living in a world filled with hatred. The AI chatter bot released by Twitter in 2016, Tay, was trained to be a Nazi within 16 hours by general users. In October 2019, Korean singer Sulli committed suicide after having been inundated with malicious online comments for months criticising her “outrageous” behaviour in the public, for example, not wearing brassiere. Not to mention the anger, anxiety and frustration that are exhibited everyday on our major news websites. However, we cannot get used to it and let ourselves fall into the Joker’s Gotham.

This project is based on the database of DOCUMENTING HATE ( that, collaborated with Google News Lab, lists media reports about hate crimes and bias incidents. The news titles are tagged on slowly rotating pyramidal particles which altogether form a constantly expanding and shrinking sphere. By clicking on each tag, you will land on the actual news page. The visual information is to unsettle the users, so is the news. However, users are encouraged to collect and submit happy news to balance the initial hatred. The comforting information will gradually replace the passive contents and change the visual from red to blue. The love/hate ratio is displayed on the bottom left.

Technical details This #JavaScript project is based on Three.js + offline datasets, with support of node.js/npm modules: canvasSketch + papaParse.

Interaction starts here :)

Screenshot of web app 網頁截屏


我們似乎生活在一個充滿仇恨的世界里:Tay,一個由 Twitter 在 2016 年發布的AI聊天機器人,在發布 16 小時內被大眾訓練為納粹。2019年10月,韓國歌手雪莉在被數月的在線惡意評論徹底淹沒後,選擇自殺。她被批評的在公眾場合"令人髮指"的行為有,諸如不穿束胸內衣。憤怒,焦慮和挫折,每天顯示在我們的主流新聞網站。然而,我們不能習慣它,讓自己落入到小丑的高譚。

該專案基於「記錄仇恨」資料庫(其與谷歌新聞實驗室合作,記錄媒體報導仇恨犯罪和偏見事件。新聞標題標記在一個不斷膨脹收縮並緩慢旋轉的球體的粒子群上。按下標籤,您將被指引向對應的新聞頁。因為對象是令人不安的新聞,視覺風格也傾向於尖銳濃烈。最初的仇恨新聞標題在鼠標停留時顯示為紅色,但是我們鼓勵使用者收集和提交好消息以平衡 - 用安慰的資訊取代負面消息,視覺效果逐漸由藍轉紅,愛恨也是。比值顯示在左下角。


此JavaScript項目基於Three.js和離線數據庫,並由node. js / npm 組件:canvasSketch + papaParse提供支持。互動從這裡開始 :)


Related posts :